🚀 go-pugleaf

Well, I think the solar cell and large electrolytic capacitor idea is a good
one ... however, if you really want a reliable system, use a radioactive-decay
thermal unit like those used on deep-space probes (e.g., Voyager, Galileo.)
The biggest drawback to these units is their expense - in the tens of thousands
of dollars. However, they are rugged, and have lives in thousands of years.
Also, they put out current in 500 mA-2 A range, enough to run a *hefty*
transmitter.

If you pick a spot that has relatively constant (or at least predictably
cyclic) winds, a small windmill + a solar panel to charge the energy-storage
system would do the trick. Or, you could play games with thermocouples if you
have a constant temperature gradient available to you.

As to energy storage, I seem to recall a "high-farad" capacitor that one of
the oil comapanies put out (Sohio?) - it acts almost like a battery, but it
isn't: it provides current all the way down to zero volts, unlike a battery.
They are available in 1-10 *Farad* denominations; however they might suffer
from the same lifetime problems as 'lytics, though.

Now, as to the *system* itself, my suggestion is to use a very-low power 8-bit
microcontroller and EEPROM: you measure off one "day" by using a timer
interrupt and a soft counter, then you update the 16-bit "day" in EEPROM.
(Come to think of it, there may be uC's out there with EEPROM *on-chip*.) To
save power and EEPROM erasure cycles (most rate their parts at about 1000 to
10,000 cycles), one could program the uC to zap the EEPROM with the timer-state
only when the power sensor says that the supply is almost exhausted.

When the day counter finally rolls over, the uC would instead run the attention
transmitter until the power is discharged each time the power sensor wakes it
up. From there on out, it transmits if there's power, rather than counting.

Interesting problem.


-Neal